540
Macromolecules 1994,27, 540-544
Raman Spectroscopic Study of Hydrogen Bonding of Polyacrylamide in Heavy Water? Naoki Tanaka Department of Polymer Science and Engineering, Kyoto Institute of Technology, Kyoto 606, Japan Kensaku Ito and Hiromi Kitano' Department of Chemical and Biochemical Engineering, Toyama University, Toyama 930, Japan Received July 16, 1993; Revised Manuscript Received October 15, 199P ABSTRACT: The Raman spectra of DzO solutions of deuterated polyacrylamide,acetamide,propionamide, and butyramide were measured to investigate the self-association of amide groups. Factor analysis and the deconvolution technique were used for the analyses of Raman spectra of these compounds, and the C=O stretching band of the amide groups could be separated into three major peaks. One of the peaks corresponded to a nonassociated amide group, and the others were assigned to different types of self-associated amide groups. Nonlinear least-squares curve fitting suggested the presence of intramolecular hydrogen bonding within a coil of polyacrylamide in aqueous solutions. Introduction Hydrogen bonding plays a crucial role in determining the properties of biopolymers. Therefore, investigation of the hydrogen bonding of water-soluble polymers (as model compounds of biopolymers) in aqueous solutions provides a basic knowledge about the properties of biopolymers. In previous studies, we used Raman spectroscopy to investigate the hydrogen bonding of poly(acrylic acid)' and poly(vinylpyrrolidinone)2 in aqueous solutions. Poly(acry1ic acid) is known to have the ability to form cross-links by hydrogen b ~ n d i n g , ~and - ~ the dimerization constant KDof the carboxylic group of poly(acrylic acid) in aqueous solutions could be estimated.* Poly(vinylpyrro1idinone) was, on the other hand, found to be hydrated in a specific way in aqueous solutionsa2In the present study we investigated the hydrogen bonding of polyacrylamide in aqueous solutions. There have been some studies on the vibrational spectrum of polyacrylamide (POAMhW Gupta and Bansi1,G for example, suggested that this polymer has a capability to form intra- or intermolecular hydrogen bonds between the amide groups in the solid state. Rheological studies of polyacrylamide in aqueous s o l ~ t i o n ~suggested -'~ the formation of hydrogen bonds between the amide groups in polyacrylamide. To study the hydrogen bonding of POAM in detail, we analyze here Raman spectra of the C=O stretching mode of N-deuterated polyacrylamide (POAM-d2) and its monomer analogues (acetamide ACAM-dz),propionamide (PRAM-&), and butyramide (BTAM-d2)) in DzO solution by various techniques. We chose heavy water, not light water, as the solvent, because the bending band of H2O overlaps with the C=O stretching band, which is unfavorable for precise analyses. A number of studies on the vibrational spectrum of amide groups in monomer analogues of POAM have been carried out.lW1 Kobayashiet al., for example, investigated hydration of amide groups in aqueous solutions by * To whom correspondence should be addressed. 'Presented at the 40th Polymer Symposium of the Society of Polymer Science, Japan, at the Okayama University of Science, Okayama, in October 1991. e Abstract published in Advance ACS Abstracts, December 1, 1993.
0024-9297/94/2227-0540$04.5010
measuring infrared spectra of the amides in dioxane-Dz0 and dioxane-Hz0 They resolved absorption spectra of the C=O stretching band into five Gaussian or Gaussian-Lorentz hybrid functions (Voigt functions) and attributed them to five types of amides with different hydration states. The C=O stretching bands in the Raman spectrum of POAM and its monomer analogues were thought to be the overlaps of several peaks which could be attributed to different hydration states of the amide groups. We used factor analysis, deconvolution, and leastsquares curve fitting to separate the measured spectrum into individual peaks. Factor analysis is a method for estimating the number of individual components in the ~ p e c t r u m . ~Recently, ~ - ~ ~ this technique was employed for the analysis of vibrational spectra of acetic a ~ i d , 3 ~ - ~ ~ i s ~ m o l y b d e n e ,p~h~~ s p h a t e ,protein,39 ~~ and polymer blends.40 The deconvolution technique, which eliminates the distortion by the slit function from the spectroscopic data,42p43has been applied in the analysis of vibrational spectroscopy.4P49 The number of hydration states of amides were estimated by this technique, and the measured spectrum was separated with least-squares curve fitting using the Marquardt method.41 By using these techniques, we analyzed the association phenomenon of polyacrylamide and its monomer analogues in DzO solutions. Experimental Section Materials. POAM-d2 of M, = 1500 (POAM1500) was obtained by deuteration of polyacrylamide (M,= 1500) from Polysciences, Warrington, PA. The polymer had been purified
by ultrafiltration beforehand. POAM-dis of M, = 1OOOOO (POAM100000)andM, = 700000(POAM700000)were prepared by fractionation and subsequent deutration of a commercial samplefrom Wako Pure Chemicals,Osaka,Japan. The molecular weights of these polymers were determined by GPC. ACAM and PRAM were purchased from Nacalai Tesque, Kyoto, and BTAM was supplied by Tokyo Kasei Co., Tokyo. These three samples were recrystallized from chloroform and deuterated to ACAM-d2, PRAM-d2, and BTAM-d2. Heavy water was obtained from CEA, Gif-sur-Yvette, France, and used for preparation of sample solutions and deuteration. Spectroscopic Measurements. Raman spectra were recorded using an NR-1100 Raman spectrophotometer (Japan Spectroscopic Co., Tokyo, Japan) with a resolution of 5 cm-l. The 488-nm line of an argon ion laser (GLG 3200, NEC, Tokyo) @ 1994
American Chemical Society
Hydrogen Bonding of Polyacrylamide in DzO 641
Macromolecules, Vol. 27, No. 2, 1994
I
I630
810 934
1 RE0
-1600
1200
1L60
1
A.
1423
\: [anis I
1000
1700
8'0 0
Wavenumber (cm-')
Figure 1. Raman spectrum of 70 wt % DzO solution of acetamide at 25 O C in the region from 700 to 1800 cm-*.
at 300-500 mW was chosen as the excitation light. The depolarization ratio was measured with a system consisting of a half-wave plate, lens, and a polarizer. All measurements were carried out in a thermostated chamber controled at 25 f 0.5 "C using a Peltier device (Model RT-IC, Japan SpectroscopicCo.). Computations. Factor analysis of the spectral data was performed with the program FCCPM written in C language by Drs. H. Kihara and T. Ozeki, Hyogo University of Teacher Education, Hyogo Prefecture, Japan. A deconvolutionprogram employing the Gauss-Seidel method was used for the analyses (the program was written by Y. Senga'2 in BASIC). A deconvolution function in this analysis has a trianglular shape with a half-width of 5 cm-l, which is equal to the spectral slit width of the present measurements. The spectra were resolved with the data decomposition program using the Marquardt method described previ~usly.~~ All calculations were performed by an NEC 9801 RA microcomputer.
1600
1500
I
1408
Wavenumber (cm-')
Figure 2. Raman spectrum of 70wt % DzO solutionof acetamide at 25 "C in the region from 1350 to 1800 cm-l: (A) isotropic spectrum; (B) anisotropic spectrum.
B. Polarization Measurements. Raman spectra of ACAM-d2 in D2O solution under parallel (Ill)and perpendicular (II)polarizations were measured by means of an analyzer in front of the slit and a polarization rotator for the excitation radiation using 90' geometry. The isotropic and anisotropic spectra were calculated using the equations and
Imi,= 4/31,
Figure 2 shows the isotropic and anisotropic spectra of the carbonyl stretching mode for a DzO solution of ACAMdz. The figure shows that the frequencies of the peak tops of the isotropic and anisotropic spectra are different from Rssults and Discussion each other, which means that there should be at least two A. Raman Spectra of Monomer Analogues. Figure components which have different depolarization ratios in 1 shows the Raman spectrum of a 70 w t 96 DzO solution the carbonyl stretching mode. One is almost at the same ~ of ACA-d2 at 25 OC. According to previous ~ t u d i e s , ~ ~ ' frequency of 1631 cm-l, and the other is at a higher the band at 1638cm-l corresponds to the C-0 stretching frequency than the former band. The same was true for band, and the complex bands in the region 1350-1480 cm-l the spectra in higher concentration regions and Dz0 are composed of the C-N stretching band and the CH3 solutions of PRAM-&. deformation bands. The bands at 810,934,and 1034cm-l C. Deconvolution. Deconvolution is a procedure for are assigned to the C-C stretching band, the NDz rocking obtaining an essential spectrum from the observed specband, and the CH3 rocking band, respectively. The C-C trum. The observed spectrum is a convolution from the stretching mode at 810 cm-1 was used as an internal original signal spectrum and the slit function of the standard. The advantages of using an internal standard spectrophotometer. The slit function affects information were described elsewhere.% We believe that the use of about the resolution, peak position, and intensity in the the C-C stretching mode at 810 cm-1 as the internal observed spectrum. In the deconvolution procedure, a standard is valid, because the intensities of the normalized slit function was eliminated from the observed spectrum, peaks of the CH3 rocking band at 1034cm-1 did not change and the original signal spectrum is presumed. The Gausswith the concentration. Seidel method is a procedure which isolates the essential The C=O stretching band exhibited a remarkable spectrum vector x by repeating calculations of the equation change with the concentration of ACAM in DzO solutions. y=Hx Previously, Kobayashi et al. investigated the hydration of amides in aqueous and DzO solutions by measuring to fit the observed spectrum vector y. In this equation, H is the slit function matrix.43 The row of H is the slit infrared spectra of amides in dioxane-Dz0 and dioxanefunction of the spectroscope. In this case, a function of H2O mixtures. They found that the absorption due to the C=O stretching vibration showed a remarkable red shift a triangular shape with a half-width of 5 cm-l (equal to accompanied by a characteristic change of the band shape the spectral slit width) was adopted as the slit function. as the water content in the medium was increased. They Figure 3 shows Raman spectra of DzO solutions of ACAMattributed this phenomenon to the change of the hydration d2 (70 w t 96) before and after deconvolution. The spectra state a t the carbonyl oxygen in equilibrium, which depends after the deconvolution showed the existence of shoulders which had frequencies below and above 1631 cm-l. We on the water content. On the other hand, the amides selffound the same shoulder in the deconvoluted spectrum of associate through the hydrogen bonding between an oxygen PRAM-& in D2O solution (data are not shown). atom of the carbonyl group and a hydrogen atom of the From the results of the polarization measurements and amino group in an argon matrix.= Kobayashi et al. also deconvolution, we conclude that the carbonyl stretching reported that the amides self-associatein dioxane solutions modes of DzO solutions of ACAM-dz and PRAM-& are at high concentrations.20~23
Macromolecules, Vol. 27, No. 2, 1994
542 Tanaka et al. I
/ A
n
1 I
Wavenumber (cm-l)
, , I
1700
160; tfaie*t,rcer
B
:--'I
Figure 3. Deconvoluted Raman spectra of the carbonyl stretching band of DzO solution of acetamide in 70 w t % DzO solution: solid line,original spectrum;broken line, deconvoluted spectrum. Table 1. Factor Analynis of D20 Solution of Amides. m 1 2 3 4 5 6 l
ACA 2276220.0 4969.5 1252.4 510.2 335.4 220.8 0.1
eigenvalue X PRA 2273930.0 7877.0 2362.1 344.8 256.8 142.4 112.6
POAM1500 2072620.0 15648.4 3217.3 1828.2 1273.0 1146.3
a Seven solutions (10-70 w t % ) for ACA and PRA and six solutions ( 1 W O wt %) for POAM were obtained by the digitization at 201 data points.
Wavenumber (cml)
Figure 4. Separation of the carbonyl stretching band of acetamide in D20 solution into Voigt components: (A) 10 w t %;
(B)70 wt %.
Among the seven values of X (six values for POAM-&), convoluted by a t least three components located around three values are by far larger than the other Xvaluea. Hence, 1610,1630, and 1650 cm-I. the number of components was concluded to be three for D. Factor Analysis. Factor analysis is a mathematical all these solutions. This is consistent with the result from technique for determining the number of linearly dethe polarization measurements and deconvolution. pendent components in a composite system. The details of the calculations can be found in the l i t e r a t ~ r e . We ~ ~ - ~ ~ E. Least-Squares Curve Fitting. According to the results from the polarization measurements, deconvoluapplied this method to seven Raman spectra of D2O tion, and factor analysis, we divided the stretching C=O solutions of ACAM-d2 and PRAM-& in the concentration band of ACAM-dz, PRAM-&, and POAM-& into three range 10-70 wt % and to six Raman spectra of a D20 individual peaks given by a linear combination of Gaussian solution of POAM-dz in the concentration range 10-60 wt and Lorentzian functions. A resolved spectrum in the 5%. The spectra were digitized a t 201 data points between C=O stretching region for the 10 wt % D20 solution of 1550 and 1750 cm-l and normalized with the C-C stretching ACAM is shown in Figure 4A. In the concentrated region band. In this region, the spectrum includes only the C=O (70 wt %), the intensity ratio of three peaks varied as stretching band. The data sets are expressed by the data shown in Figure 4B. The intensity of the central peak matrix D,which is the product of spectrum matrix R and decreased, whereas those at both ends increased. The component matrix C. same was found for the peaks in the C=O stretchingregion D=RC (1) for the D2O solutions of PRAM-d2 and POAM-& In essence,we diagonalize the covariancematrix Z which Self-associationof the amides through hydrogenbonding is obtained using eq 2 by the Jacobi method and get the might be remarkable under the present experimental eigenmatrix Q and eigenvector E. conditions, because the mole fraction of the amides was relatively high (0.0354.433 for ACAM-d2 and 0.029-0.383 Z = 'DD = QE'Q (2) for PRAM-d2) and the number of water molecules which The Z matrix is also expressed as surrounded the amide molecules was not sufficient to prevent the amides from self-associating. Two kinds of Z = 'DD = t(RC)(RC) = 'C('RR)C (3) self-association were proposed for alkyl amides. One is a Comparing eqs 2 and 3, we obtain linear hydrogen bonding between a hydrogen atom of an NH2 group and an oxygen atom of a carbonyl group of 'Q = C, ('RR) = E (4) another molecule. The other was a closed dimer formed by two hydrogen bonds like carboxylic acids.60d2 Since The resulting eigenvalue X in eigenmatrix Q represents bands A and C increased their areas with concentration a relative importance in the relevancy of the associated (Figure 4A,B),these two bands correspondto the stretching eigenvector; the number of larger eigenvalues corresponds to the number of independent components. Table 1shows of the carbonyl group of the associated amide. Band B X values against m, the number of factors assumed for corresponds to the C=O stretching mode of a non-selfACAM-&, PRAM-d2, and POAM-d2 in DzO solution. associated amide molecule.
Macromolecules, Vol. 27, No. 2, 1994
Hydrogen Bonding of Polyacrylamide in D2O 543
1638
\
1631
I
1630
1622 /
Wavenumber (cm-l)
lB
A
'634
Figure 6. Raman spectra of DzO solution of polyacrylamide (Mw = 1500) at 25 "C in the region from 1600 to 1700 cm-1 (concentration range 10 (B)to 60 wt 5'% (A)).
Wavenumber (cm-')
Figure 5. (A) Raman spectra of DzO solution of acetamide at 25 "C in the region from 1600 to 1700 cm-l (concentrationrange 10 (B)to 70 w t % (A)). (B)Raman spectra of D20 solution of propionamide at 25 "C in the region from 1600 to 1700 cm-l (concentration range 10 (B)to 70 w t 7'% (A)). According Kobayashi et al.,20-23the hydration at the carbonyl oxygen is in an equilibrium state among the various states of hydration depending on the number of hydrating water molecules. The C=O stretching modes of the different hydration states were separated into five Gaussian or Voigt functions. Their result is contradictory to our results, which were evaluated by the factor analysis, which indicated the predominance of the three components. The other two components found by Kobayashi et al. did not appear, probably because they were too small compared to the two components of the self-association. F. Self-Association of Alkyl Amides. Figure 5 shows the comparison of the spectral change of the C=O stretching bands of ACAM-dz and PRAM-dz in D2O solutions. Both of these bands similarly changed their shapes as the concentration varied. The half-width of the carbonyl stretching band of PRAM was, however, always larger than that of ACAM at the same Concentration. The curve resolution clarified that the areas of band A and band C for PRAM-d2 in Figure 5B are larger than those for ACAM-d2 in Figure 5A. This shows that the selfassociation of PRAM in D2O solution is more significant than that of ACAM. The same procedure was used for BTAM-d2 in D2O solutions at 10 and 20 wt % . BTAM-d2 could not be dissolved above ca. 30 wt 5%. If hydrophobicinteraction influences the self-association of alkyl amide as found in the self-association of carboxylic acid in aqueous solutions,m62BTAM-d2 must self-associate more strongly than PRAM-d2. The spectra of the C 4 stretching band of BTAM-dzat these concentrations were, however, almost identical to that of PRAM-dz at the same concentration. Therefore, the quantitative differ-
C Figure 7. Schematic diagram showing three types of hydrogen bonding expected in Dz0 solution of polyacrylamide: (A) intramolecular hydrogen bonding in the helical structure; (B) intramolecular hydrogen bonding in extended segments; (C) intermolecular hydrogen bonding. ence between the self-associations of ACAM-dz and PRAM-dz may be due to the difference in their structures. G. Association between Amide Groups of POAM. By the similar method adopted for ACAM-d2 and PRAMdz, we investigated the C=O stretching band of POAMdz in D2O solutions. As shown in Figure 6, the C-0 stretching band of POAM1500 shifted to a higher frequency as the concentration was increased. These bands are also composed of three individual peaks like ACAMd2 and PRAM-d2 as clarified by the factor analysis (Table 1). The curve fitting of this band by the Voigt function showed that areas of the bands which were assigned to the associated C=O stretching band were larger than those of PRAM-dz. Therefore,the association of an amide group with another amide was enhanced in the order of ACAMdz,PRAM-dz,and POAM-d2. POAM-d2 can have two kinds of cross-linkings: intraand intermolecular cross-linkingsthrough hydrogen bonding. We compared the C=O stretching bands of POAM1500, POAM100000,and POAM700000 in D20 solutions of 10 wt 5%. The spectra of these solutions were almost the same, which means that the molecular weight does not affect the cross-linking. Susceptible forms of the cross-linking of POAM-dz are schematically presented in Figure 7. The helical structure of type A and the extended structure of type B show intramolecular cross-linking, and type C is due to intermolecular cross-linking. Polymer complexes through hydrogen bonding (such as poly(acry1ic acid)-polyacryl-
Macromolecules, Vol. 27, No. 2, 1994
544 Tanaka et al.
amide, poly(acry1ic acid)-poly(vinylpyrrolidinone), and poly(acry1ic acid)-poly(ethy1ene glycol)) are formed by the cooperative interaction between two macromolecules, and drastic chain-length dependences are found in this p h e n ~ m e n o n .If~ we ~ assume cross-linking of type C for the association of POAM-dz in D2O solutions, the association should increase with molecular weight due to the cooperative interaction between two macromolecules. As mentioned above, however, the degree of intramolecular hydrogen bonding of POAM1500 in D20 solutions was the same as those of higher molecular weight samples which can easily form the helical structure. The degree of polymerization of POAM1500 is ca. 20, and the polymer of such a small degree of polymerization cannot form the helical structure of type A, which shows that the hydrogen bond of type B is relatively more plausible for POAM. The molecular weight dependence of the hydrogen bonding of poly(acry1ic acid) is the same as that of POAM: the KD values (dimerization constant for carboxylic groups) of poly(acry1ic acid) of M , = 1000 and that of M , = 90000 were almost the same.l The hydrogen bond in the extended structure may be formed for both poly(acry1ic acid) and POAM in aqueous solutions. In the case of poly(acry1ic acid), the amount of hydrogen bonding of the polymer was smaller than that of the monomer analogue,’ whereas the reverse was found for the hydrogen bonding of POAM. Most of the hydrogen bonds between carboxyl groups of poly(acry1ic acid)s in aqueous solution are of a %lased" type in which two carboxyl groups form a ring structure through the two hydrogen bondings between oxygen atoms of carbonyl groups and hydrogen atoms of hydroxyl groups.’ The amount of the hydrogen bonding of poly(acry1ic acid) is, therefore, smaller than that of its monomer analogue because of a steric restriction. “Closed”type and “linear” type are, on the other hand, proposed for the forms of the hydrogen bonding of POAM in D20 solutions. The steric restriction for the hydrogen bonding is small in this case, which may be the reason for the difference between the hydrogenbondings ofpoly(acry1icacid) and that of POAM.
Acknowledgment. We wish to thank Professor N. Ise, Department of Polymer Chemistry, Kyoto University, for his continuing interest, encouragement, and help. This work was supported by the Ministry of Education, Science, and Culture, Japan (Grant-in-Aid 03453112). References and Notes (1) Tanaka, N.; Kitano, H.; Ise, N. Macromolecules 1991,24,3017. (2) Tanaka, N.; Ito, K.; Kitano, H. Makromol. Chem. Submitted. (3) Priel, Z.; Silberberg, J. J.Polym. Sci., Part A-2 1970,8,689,705, 713. (4) Miyamoto, T.; Cantow, H. J . Makromol. Chem. 1973,169,211. (5) Koenig, J. L.; Angood, A. C.; Semen, J.; Landon, J. B. J. Am. Chem. SOC.1969,91,7250. (6) Gupta, M. K.; Bansil, R. J . Polym. Sci.,Polym. Phys. Ed. 1981, 19, 353.
(7) Kulicke, W. M.; Siesler, H. W. J . Polym. Sei., Polym. Phys. Ed. 1982, 20, 553.
(8) Suh, J. S.; Michaelian, K. H. J . RamanSpectrosc. 1987,18,409. (9) Bovey, F. A.; Tiers, G. V. D. J. Polym. Sci., Part A 1963,1,849. (10) Schurz, J. J . Polym. Sci., Polym. Lett. Ed. 1984,22,43. (11) Kulicke, W.-M.; Kniewake, R. Makrcmol. Chem. 1980,181,823. (12) Kulicke, W.-M.; Kniewske, R.; Klein, J. Prog. Polym. Sci. 1982, 8, 373. (13) Suzuki, I. Bull. Chem. SOC.Jpn. 1962, 35, 1279. (14) Uno, T.; Machida, K.; Saito, T. Bull. Chem. SOC.Jpn. 1969,42, 897. (15) Uno, T.; Machida, K. Spectrochim. Acta 1971,27A, 833. (16) Lifson, S.; Hagler, A. T.; Dauber, P. J . Am. Chem. SOC. 1979, 101, 5111. (17) Fogarasi, G.; Balazs, A. J . Mol. Struct. 1985,133, 105.
(18) Kuroda. Y.: Saito. Y.: Machida. K.: Uno. T. Bull. Chem. SOC. Jpn. 1972, #, 2371. ’ (19) Usha Bai, P.; Ramiah, V. Indian J . Pure Appl. Chem. 1974,12, .
I
I
1A X
(20)Kibayashi, M.; Niehioka, K. J. Phys. Chem. 1987,91, 1247. (21) Kobayashi, M.; Kobayashi, M. J. Phys. Chem. 1980,84, 781. (22) Kobayashi, M. Kyoto Daigaku Genshiro Jikkensho 1986, KURRZ-TRZ75 Suiyouekikei no Characterization 56. (23) Kobayashi, M. Kyoto Daigaku Genshiro Jikkensho Gakujutsukouenkai Kouen Youshishu 1987,21,73. (24) Kutezelnigg, W.; Mecke, R. Spectrochim. Acta 1962, 18, 549. (25) Rasanen, M.; Muruto, M.; Kivinen, A. Ber. Bunsen-Ges. Phys. Chem. 1978,82, 63. (26) Kydd, R. A.; Dunham, A. R. C . J . Mol. Struct. 1980, 69, 79. (27) Bukowska, J. Pol. J. Chem. 1984,58, 243. (28) Maithis, R.; Bouissou, T.; Ayed, N.; Baccar, B. Spectrochim. Acta 1986,42A, 519. (29) Adachi, A.; Takeuchi, H.; Tasumi, M. Koen Yoshishu Bunshi Kozo Toronkai; The Chemical Society of Japan, Tokyo 1979, 96.
(30) Bala, S. A.; Panja, P.; Ghosh, P. N. J . Mol. Struct. 1987, 157, 339. (31) Bohn, B. B.; Andrews, L. J . Phys. Chem. 1989,93, 5684. (32) Malinowski, E. R.; Howery, D. G. Factor Analysis in Chemistry; Wiley: New York, 1980. (33) Ozeki, T.; Kihara, H.; Hikime, S. Bunseki Kagaku 1986,35, 1986. (34) Ozeki, T. Bunseki 1991, 252. (35) Bulmer, J. T.; Shurvell, H. F. J . Phys. Chem. 1973, 77, 256. (36) Ng, J. B.; Shurvell, H. F. J . Phys. Chem. 1987,91,496. (37) Ozeki, T.; Kihara, H.; Hikime, S.; Ikeda, S. Anal. Sci. 1987,3, 285. (38) Ozeki, T.; Kihara, H.; Hikime, S. Anul. Chem. 1987,59, 945. (39) Lee,D.C.;Haris,P. I.;Chapman,D.;Mitchell,R.C.Biochemistry 1990,29,9185. (40) Rao, G. R.; Caatiglioni, C.; Gusaoni, M.; Zerbi, G.; Martwcelli, E. Polymer 1985,26,811. (41) Marquardt, D. W. J. SOC. Ind. Appl. Math. 1963,11,431. (42) Kawata, S.; Minami, S. Bunkoukenkyu 1985,35,24. (43) Minami, S. Kagakukeisoku no tame no hakeidetashori CQ Shuppan, Tokyo, 1986. (44) Kauppinen, J. K.; Moffat, D. J.; Mantach, H. H.; Cameron, D. C. Appl. Spectrosc. 1981, 35, 271. (45) Kauppinen, J. K.; Moffat, D. J.; Mantach, H. H.; Cameron, D. C. Appl. Spectrosc. 1983,37,67. (46) Jones, R. N.; Shimokoahi, K. Appl. Spectrosc. 1983,37,59. (47) Shimokoshi, K.; Jones, R. N. Appl. Spectrosc. 1983,37,67. (48) Kallanjiev, T.; Petrov, V. P.; Georgiev, G.; Miteva, M.; Nickolov, Zh. J. Mol. Struct. 1984, 115, 409. (49) Zhelyaskov, V.; Georgiev, G.; Nickolov, Zh. J. Raman Spectrosc. 1988, 19, 405. (50) Tanaka, N.; Kitano, H.; he, N. J . Phys. Chem. 1991,96,1603. (51) Tanaka, N.; Kitano, H.; Ise, N. J. Phys. Chem. 1990,94,6290. ( 5 2 ) Ohno, H.; Abe, K.; Tsuchida, E. Makromol. Chem. 1978,179, 755. (53) Osada, Y. J . Polym. Sci., Polym. Chem. Ed. 1979, 17, 3486.